Materials Letters 235 (2019) 193–196

Contents lists available at ScienceDirect

Materials Letters
journal homepage: www.elsevier.com/locate/mlblue

Effect of prestrain on hydrogen diffusion and trapping in structural steel

Cheolho Park a, Namhyun Kang a,⇑, Myunghyun Kim b, Stephen Liu c
a
Department of Materials Science and Engineering, Pusan National University, Busan 46241, Republic of Korea
b
Department of Naval Architecture and Ocean Engineering, Pusan National University, Busan 46241, Republic of Korea
c
Department of Metallurgical and Materials Engineering, Colorado School of Mines, CO 80401, USA

a r t i c l e i n f o a b s t r a c t

Article history: Hydrogen permeation tests were electrochemically performed to measure hydrogen diffusion and trap-
Received 19 September 2018 ping in EH 36 steel. Permeation transients were measured through charging and decaying cycles, for
Received in revised form 5 October 2018 specimens prestrained to varying degrees (0–20%). With increasing prestrain, the effective hydrogen dif-
Accepted 7 October 2018
fusivity (Deff) decreased, while the total/reversible hydrogen content (Co and Cr, respectively) increased.
Available online 8 October 2018
The immobile dislocation of the 20%-prestrained specimens decreased Deff by approximately ten times
and increased Co and Cr by approximately three times. The amount of irreversibly trapped hydrogen
Keywords:
(Cirr) remained approximately the same, regardless of the prestrain. The dislocations delay the motion
Steel
Electrochemical permeation
of hydrogen and contain more reversible hydrogen.
Prestrain Ó 2018 Elsevier B.V. All rights reserved.
Hydrogen embrittlement
Hydrogen diffusion
Hydrogen trapping

1. Introduction increase the strength. However, the influence of dislocations on HE

Hydrogen provides purely clean energy in current times. To use
hydrogen as an energy source, it is necessary to build such a facility
such as a hydrogen carrier and recharging station. Hydrogen
embrittlement (HE) should be studied for structural steels to con-
struct them. Hydrogen atoms in steel can be diffused and trapped.
The trapping in defects determines the hydrogen distribution in
the microstructure, and HE is correlated with hydrogen-defect
interaction. An irreversible trap is associated with a site having
a large activation energy, such as phase boundaries and
inclusion/precipitate interfaces [1–4]. Hydrogen atoms trapped in
these sites are considered as non-diffusible hydrogen. However,
reversible trap sites, such as grain boundaries and dislocations,
have a lower activation energy. Hydrogen atoms in reversible traps
are diffusible, and these traps influence the effective hydrogen
diffusivity (Deff) and HE resistance.
HE is associated with the alloy composition, strain rate, and dif-
fusible hydrogen content [5–8]. Particularly, HE depends on the
microstructural features such as carbide type and shape, grain size,
and dislocation density [9–11]. Most structural steels are typically
fabricated by thermomechanical-controlled processing (TMCP),
yielding a large density of dislocations and nanocarbides to thereby
⇑ Corresponding author.
E-mail address: nhkang@pusan.ac.kr (N. Kang).
is disputed [12,13]. In this study, the effects of dislocation density
on hydrogen diffusion and trapping in EH 36 steel were investi-
gated using hydrogen permeation tests, while maintaining the
chemical composition, phase type, and grain size of the steel.
2. Experimental procedure
The study used EH 36 steel composed primarily of Fe, 0.08 wt%
C, 0.31 wt% Si, and 1.54 wt% Mn. EH 36 steels typically have a com-
plicated microstructure. In order to identify purely the prestrain
effect on HE. The steels were reheated to 1150 K and held at this
temperature for 12 hr. Prestrains of 0–20% were applied to vary
the dislocation density in the tensile specimen, using uniaxial ten-
sile tests. Tensile specimens of gauge length 25 mm were prepared
according to the ASTM A370-03 standard.
A two-component electrolytic permeation cell was designed
following ISO 17081. The EH 36 membranes between the anodic
and cathodic sides were ground with SiC up to #2400 grit, to obtain
a final thickness of 500 ± 5 lm, and an exposed surface area of
314 mm2. The hydrogen permeation test was carried out at
300 K. The cathodic sides of the specimens were electrochemically
charged with a 0.1 N NaOH solution, at a constant current density
of 5 A m
2. The anodic side was potentiostatically maintained at
+250 mV, with a 0.1 N NaOH solution, relative to a saturated
calomel electrode.

https://doi.org/10.1016/j.matlet.2018.10.049
0167-577X/Ó 2018 Elsevier B.V. All rights reserved.
194 C. Park et al. / Materials Letters 235 (2019) 193–196

3. Results and discussion
3.1. Microstructural behavior with respect to prestrain
Fig. 1a shows the tensile stress–strain curves for annealed EH
36 specimens. The yield and ultimate tensile strength (UTS) were
280 and 420 MPa, respectively, and the total elongation was 52%.
The strain value for the UTS was 28%. Regardless of the amount
of prestrain, phase maps in Fig. 1b, d, and f exhibited the same
grain size, at 25 ± 1 lm. A dual phase structure of ferrite
(94.5 ± 0.5%) and cementite (5.5 ± 0.5%), indicated with red and
green colors, can be seen.
Fig. 1c, e, and g show kernel average misorientation (KAM)
maps. The blue-colored area in the maps represents the least
deformed area, and the smaller KAM value, and the red-colored
area represents the most deformed area, and the larger KAM value.
The KAM value was 0.69° for the 0% prestrained specimen (Fig. 1c),
which increased to 1.91° for the 20% prestrained specimen
(Fig. 1g). Large KAM is known to be associated with a large disloca-
tion density [14], indicating that by applying prestrain to the
annealed EH 36 specimens, this study produced specimens with
varied dislocation densities.
3.2. Electrochemical permeation behavior for repeated charging and
decay
Fig. 2a shows the current density behavior on the anodic side
for the 10% prestrained specimen. Initially, the permeation rate
increased significantly, with current application, reaching the
steady-state current density (Iss), followed by significant decay
when the current was disconnected. The second and third charging
rounds started when the anodic current fell below 0.3 lA, and the
same pattern of permeation behavior was reproduced, regardless
of the prestrain amount.
For the permeation transient, the total hydrogen concentration
(Co) was calculated as the sum of hydrogen in the lattice, and
trapped in reversible (Cr), and irreversible sites (Cirr). Fig. 2b shows
a direct comparison of the area between the 1st, 2nd, and 3rd per-
meation curves of the 10% prestrain steels. In the steady state, both
reversible and irreversible traps reached their equilibrium condi-
tions. Reversible and lattice hydrogen were released during the
decaying phase, indicating that the irreversible traps were satu-
rated and that the reversible traps with lattice diffusion were only
effective for the 2nd and 3rd permeations. On this basis, the area
between the 1st and 2nd permeation curves can provide an

Fig. 1. Application of prestrain using uniaxial tensile test: (a) tensile stress–strain curves for annealed EH 36 specimens; phase and KAM maps, respectively, for (b, c) 0%, (d, e)
10%, and (f, g) 20% prestrained specimens.

Fig. 2. Anodic side current density behavior for 10% prestrained specimen: (a) 3 consecutive charge and decay cycles; (b) direct comparison of the area between the 1st, 2nd,
and 3rd permeation transients.
 C. Park et al. / Materials Letters 235 (2019) 193–196 195

approximation of irreversibly trapped hydrogen content, and the Table 1

2nd permeation transient can be used to determine the Deff. Hydrogen diffusivity and concentration for EH 36 steels using electrochemical
permeation.
The amount of Cirr was calculated using Eq. (1), and hydrogen
permeation flux at steady-state (j) was calculated using Eq. (2). Prestrain Deff (m2 s
1) C o (mol m
3) C irr (mol m
3) C r (mol m
3)
Z (%)
DQ
C irr ¼ ; DQ ¼ fjð2nd permeationÞ
jð1st permeationÞgdt  A 0 2.38E
10 46.41 6.24 33.91
FL 10 5.16E
11 79.01 6.15 59.38
ð1Þ 15 3.71E
11 108.79 6.34 81.44
20 3.14E
11 134.91 6.44 102.17

C o  Deff
j¼
L
In Eq. (1), Q is charge, F is Faraday’s constant (96,500 C mol
1), A
is the surface area exposed, and L is the thickness of the steel mem-
brane, and in Eq. (2).
Irreversible hydrogen would be captured at the interfaces
between cementite and ferrite, at the carbide interface, and in
the microvoids produced in tensile testing [15]. Cirr values in the
variously prestrained (0–20%) steels were calculated to fall in the
range of 6.14–6.42 mol m
3; thus, the values were mostly the same
regardless of the prestrain, as were the amounts of cementite
(Fig. 1b, d, and f).
3.3. Influence of a prestrain on the permeation behavior
The current density behavior depends on specimen thickness.
To be independent of specimen thickness, Fig. 3a and 3b were plot-
ted using the normalized current density (I(t)/Iss) as a function of
the time to square thickness ratio (s mm
2), during (a) charging,
and (b) decay transients, with various prestrains. As the prestrain
increased, the normalized current density increased slowly, and
the maximum current density decreased (Fig. 3a). The hydrogen
trapped by the dislocations would either pin them, move with
them, or be left behind, depending on hydrogen mobility, the
velocity of the dislocations, and the magnitude of the binding
energy. In this study, hydrogen trapping was predominant, as
mobile dislocation was normally absent in the prestrained speci-
mens, therefore delaying the Deff. Fig. 3b shows that the normalized
current density during the decay transients decreased slowly as
the prestrain increased. The slow decrease of the decay transients
for the large prestrain was consistent with its dislocation density
(Fig. 1c, e, and g). The area difference between the decay curves
of 20% and 0% prestrained specimens could be considered as the
amount of trapped hydrogen in dislocation.
The Deff was calculated with the time lag method, using Eq. (3)
in the second transient curve.
L2
Deff ¼
6  t lag
ð2Þ
In Eq. (3), L is sample thickness (m) and tlag is the time to
achieve a value of I(t)/Iss = 0.63. Co (mol m
3) was calculated using
Eq. (4).
Iss  L
Co ¼ ð4Þ
F  Deff
Table 1 summarizes the values of Cirr, Deff, Co, and Cr, calculated
by subtracting Cirr from Co. As the prestrain increased from 0 to
20%, Deff decreased, while Co and Cr increased; there was no varia-
tion in Cirr. The Deff for the 20% prestrained specimens was 10 times
slower than that for the 0% specimens, indicating that the disloca-
tion decreased the hydrogen diffusivity and increased the hydro-
gen concentration in the specimens. The large density of the
dislocations would delay the motion of hydrogen and improve
HE resistance, while the large amount of reversible hydrogen
trapped in the dislocations should be critically controlled, as it
can provide possible hydrogen sources, in the presence of triaxial
stress and deteriorate HE resistance in engineering applications.
4. Conclusions
This study evaluated the effects of prestrain on hydrogen diffu-
sivity and concentration, using hydrogen permeation testing on EH
36 steel. The major conclusions are as follows:
1) Hydrogen permeation tests successfully measured Deff, Co, Cr,
and Cirr.
2) The immobile dislocation in the 20% prestrained specimens
delayed the Deff approximately ten times, and increased Co
and Cr by approximately three times.
3) The Cirr showed mostly the same value, regardless of
prestrain.
4) The dislocations would delay the motion of hydrogen, while
the large Cr trapped in the dislocations should be critically
ð3Þ controlled.

Fig. 3. Normalized current density as a function of the time to square thickness ratio (s mm
2) during (a) charging and (b) decay transients.
196 C. Park et al. / Materials Letters 235 (2019) 193–196
Acknowledgements
This work was supported by Ministry of Trade, Industry &
Energy Grant No. 10080728, and National Research Foundation
(NRF) of Korea Grant No. 2011-0030013, funded by the Korea gov-
ernment (MSIT) through GCRC-SOP.
References
[1] X. Cheng, X. Cheng, C. Jiang, X. Zhang, Q. Wen, Mater. Lett. 213 (2018) 118–121.
[2] X. Zhou, J. Song, Mater. Lett. 196 (2017) 123–127.
[3] X. Xing, W. Chen, H. Zhang, Mater. Lett. 152 (2015) 86–89.
[4] B.G. Mytsyk, Y.L. Ivanytskyi, A.I. Balitskii, Y.P. Kost, O.M. Sakharuk, Mater. Lett.
184 (2016) 328–331.
[5] G.K. Padhy, H. Terasaki, Y.I. Komizo, Mater. Lett. 136 (2014) 22–25.
[6] H. Addach, P. Berçot, M. Rezrazi, M. Wery, Mater. Lett. 59 (2005) 1347–1351.
[7] P.W. Liu, J.K. Wu, Mater. Lett. 57 (2003) 1224–1228.
[8] A. Oudriss, S. Le Guernic, Z. Wang, B. Osman Hoch, J. Bouhattate, E. Conforto, Z.
Zhu, D.S. Li, X. Feaugas, Mater. Lett. 165 (2016) 217–222.
[9] H.J. Kang, J.S. Yoo, J.T. Park, S.T. Ahn, N. Kang, K.-M. Cho, Mater. Sci. Eng., A 543
(2012) 6–11.
[10] C. Park, N. Kang, S. Liu, Corros. Sci. 128 (2017) 33–41.
[11] J.B. Lee, N. Kang, J.T. Park, S.T. Ahn, Y. Do Park, I.D. Choi, K.R. Kim, K.M. Cho,
Mater. Chem. Phys. 129 (2011) 365–370.
[12] M. Koyama, H. Springer, S.V. Merzlikin, K. Tsuzaki, E. Akiyama, D. Raabe, Int. J.
Hydrogen Energy 39 (2014) 4634–4646.
[13] J.A. Donovan, Metall. Trans. A 7 (1976) 1677–1683.
[14] J. Albrecht, I.M. Bernstein, A.W. Thompson, Metall. Trans. A 13 (1982) 811–
820.
[15] A. Clair, M. Foucault, O. Calonne, Y. Lacroute, L. Markey, M. Salazar, V. Vignal, E.
Finot, Acta Mater. 59 (2011) 3116–3123.